DailyDirt: Faster Than A Speeding Bullet…
from the urls-we-dig-up dept
Algorithmic trading is changing the way the stock market game is played, as human reaction times are no longer fast enough to keep up with machines that can make nearly instantaneous decisions about stock trades. A human chess grandmaster takes about 650 milliseconds to recognize when a king piece has been put in check, but it doesn’t take anywhere near that long for a computer to perform a few profitable transactions. In 2010, the “flash crash” caused the stock market to plunge for a few minutes, and the SEC published a report on its findings of what happened on that day, but there may be a lot more market instability caused by machines — and we’re only started to recognize the implications. Checkmate, humans!
- Since 2006, the stock market spiked and crashed thousands of times — but no one really noticed. High frequency trading algorithms can produce “ultrafast extreme events” (UEEs) in which computers are causing millisecond-scale price fluctuations, and these trading events could be changing the way the stock market behaves on longer time scales. (But it’s hard to know exactly what’s going on….) [url]
- The NASDAQ mysteriously stopped all trading for over three hours over the summer, due to technical glitches that have raised concerns about a lack of transparency for how the market operates. Quite a bit of high frequency trading can happen over the course of a few hours…. [url]
- Latency arbitrage relies on high frequency trading and potentially makes billions of dollars every year for companies involved in high speed trading. An academic study finds that this kind of arbitrage can hurt the average investor and proposes a non-continuous market that would eliminate the advantages of high frequency trading. [url]
If you’d like to read more awesome and interesting stuff, check out this unrelated (but not entirely random!) Techdirt post via StumbleUpon.
Filed Under: algorithms, flash crash, hft, high frequency trading, latency arbitrage, stock market, ultrafast extreme events
Companies: nasdaq
Comments on “DailyDirt: Faster Than A Speeding Bullet…”
TAX WALL STREET! 1% per transaction tax.
We need to end the casino. Won’t kill them to be slightly LESS Rich, and we’d stifle unproductive gambling that puts the rest of us at risk. Wall Street doesn’t actually contribute anything to production, simply makes phony paper to GRIFT away the real profit that laborers make. — And when their bets go wrong as in 2008, why, they have gov’t tax The People to make them whole. It’s private profits and public losses. You’ll be paying off their losses and the phony “debt” for rest of your life, kids. — And it’s not new. This is same “laissez-faire” rampant speculation as the 1920’s, leading to the Great Depression. Glass-Steagall and various other regulation kept the Money Manipulators under control for decades, but again the brakes are off. — Of course you kids think that can’t happen, but it’s certain. Won’t happen all at once, but increases every day. Soon it’ll be your bank accounts taken by “bail-in” theft as in Cyprus. And forget retiring. You’ll be locked into a Foxconn hive with suicide nets.
Re: TAX WALL STREET! 1% per transaction tax.
OOTB, please do us all a favor and take your meds.
Re: TAX WALL STREET! 1% per transaction tax.
Even a broken clock is right twice a day, I guess this is ootb’s time.
I'm a pro trader
Yup, the bots are there, and sometimes they screw up mightily and I take advantage of them – human judgment still rules over an algo written by some guy who graduated but couldn’t get a real job in physics.
Check out nanex, who track this sort of thing. You can often make money off the noise the algos generate.
They’ve only discovered a couple of the basics of DSP. They know the step function. They know how to run stops, and take out a bid or ask stack to manipulate. So did open-cry traders in the day. You win by using a different attention span than they do, ride the moves they create for the middle part – greed kills, but smart makes money.
I’ve seen the algos in action – I would trust those guys to write a program or build an audio amp that wouldn’t oscillate all by itself. Hook 10 or 1000 together and the chaos is fun to watch sometimes. Not that they should be there, but you gotta live with what is if you can’t change it.
Re: I'm a pro trader
Nice try ameritrade
Is anyone else seriously questioning the reasoning behind letting people get ridiculously rich from these high-frequency trades when the decision process is ultimately handled by a computer? It’s like World of Warcraft gold-farming with real money…
Re: Re:
Frankly, no, but I speak only by myself here.
I don’t mind people getting rich, I don’t really care how much others make, what I do care is the playing field, if it is all even then we are good, if some of them start to create distortions in it at the cost of everybody else, by artificial rules, then I take notice.
With that said, what is the artificial rules that exclude everyone from doing the same? how do we get rid of them or get around them?
Re: Re: Re:
level playing field …. hahahahaha, you so funny.
No, it’s not a level playing field. Most of the money made by algos is by doing huge volumes for a fraction of a penny a share, faster than you can. To do that, takes a colo and more money than most will ever have. It’s fairly easy to see it going on with either nanex or level II trade data. They can only pull that stuff off with very liquid stocks, which I tend not to trade when I see the algo’s signatures.
They tend to “cheat” – if you can get in and out of a trade faster than your funding source can react, you might not actually have had the money. They cheat by bid/ask flooding, canceling orders so fast no one can bite – while watching to see if anyone tries and so on. They can create such a flood of such fake “order stuffing” that the exchanges barf – what amounts to a DOS.
But they are just there, and there appears to be nothing that can be done about them, other than work with a longer attention span than they do. They get fooled by fake tweets and headlines frequently, and some have mastered faking them out that way – just another form of cheating, this time by humans.
What was really hilarious was the day Knight trading accidentally put their test harness (a fake stock market designed to test algos on and be crushed by them) online instead of their new version algos, and lost 444 million in about 45 min, buying high and selling low as fast as it could. Made a few month’s income that morning, just knowing that stuff doesn’t really double or half in such a short time with no news, and using my judgment on some quick trades.
Nice try at what, AC? So brave to insult behind no name, eh? Just saying what I know and about something I do all day every day. If you have a different, knowledge/fact based take, lets hear it.
Re: Re:
There is a little clickything labeled “reply to this”, noobs may not realize its usefulness.
I think there was a Ghost in the Shell episode that revolved around this concept…
I'm just an ex-clown....
…..so feel free to ignore me, but I’d be embarrassed to admit it if I were a “pro trader.” Sounds rather meretricious, don”t ya know.
Charge a penny or fraction thereof for each order made and subsequently cancelled and HFT might get toned down enough. If not, just introduce random delays of, say, 100 to 500 milliseconds to random transactions at random times and watch the fun.
And if fake order stuffing is causing exchanges to barf, exactly how is that different from a deliberate DOS attack? Does the CFAA cover DOS? If it doesn’t, why not? If it does, why isn’t it used to deter these arrogant greedsters?
Take away the grease
This so called high frequency trading is also known as “front running”. Oh sure, GS will bore you for hours telling you how this makes the market better by adding liquidity or some other nonsense, but it is front running, pure and simple. Watch the trades, insert yours ahead of the line and profit by a fraction of a penny because they can. There is no value here to anyone. If any of us tried to front run trades, we’d be sharing a cell next to Madoff.
The solution is simple; add enough friction to trading that these low margin high frequency trades can’t possibly make money. The problem is the big trading houses’ expenses are so low; they can make profits from many tiny transactions. Eliminate that advantage and the problem will go away.
Zing!
That’s what my last girlfriend said.